58 



Fishery Bulletin 103(1) 



1.0 1 

 o 8 

 0.6 

 0.4 

 0.2- 



0.0 



30 



90 100 



30 40 50 60 70 80 



1.0-1 



30 40 50 60 



0.4 

 0.0 



B 



n 



-0.4-1 



34.5 38.5 42.5 46.5 50.5 54.5 58.5 62.5 66.5 



1.0 i 



00 



D 



~n 



Ji 



-0 5-1 — 



34 5 38 5 42 546 550 554 558 562 .566.570.574.5 



625 66.5 70.5 74 5 78 5 82.5 86 5 90 5 



CL (mm) 



Figure 4 



Fitted selectivity curves from the selected models (identified in Table 1) and their deviance 

 residuals for a range of stretched square meshes under aquarium conditions. A and B are 

 for 62-mm mesh, C and D are for 75-mm mesh and E and F are for 100-mm mesh. 



52 57 62 67 72 



CL-class mid-point (mm) 



Figure 5 



Relationship between proportion of rock lobsters retained 

 by 62-mm mesh trap and carapace length (aggregated 

 into 5-mm size classes). Filled circles represent data 

 from traps with entrance funnels sealed [Proportion 

 retained = 0.58/mmxCL (mm) + 48.35; r 2 = 0.79; n=l\ 

 P= 0.008), whereas open circles represent data from traps 

 with entrance funnels open {Proportion retained = 1.0l/ 

 mmxCL (mm)-38.94; r 2 = 0.62; n = l\ P=0.036). 



larger than 84 mm CL escaped. This resulted in contact 

 selectivity curves for which estimates of both L 50 and 

 SR decreased when captive lobsters were denied the op- 

 portunity to escape through the entrance funnels (Fig. 

 6; Table 2). This finding indicates that considerable 

 numbers of lobsters of all sizes can escape commercial 

 traps by the entrance funnels. 



Irrespective of whether the entrance funnels of the 

 traps was sealed, the symmetrical logistic model was se- 

 lected in preference to the asymmetrical Richards model 

 (Table 2), and the selected model fitted the data reason- 

 ably well (Fig. 6; Table 2), although not as well as the 

 models fitted to the aquarium data (Fig. 4; Table 1). 



In comparison with the selectivity curves from the 

 aquarium trials with 100-mm mesh, the corresponding 

 curves from field trials indicated that greater numbers 

 of larger lobsters are retained in practice than under 

 laboratory conditions (Figs. 4 and 6). This finding in- 

 dicates that some lobsters are retained in commercial 

 traps, even though they can escape, which goes some 

 way to explaining the more "scattered" fit of the logistic 

 model compared to the field data. 



